You also will want to read http://www.asktog.com/TOI/toi06KeyboardVMouse1.html and http://www.asktog.com/TOI/toi22KeyboardVMouse2.html (opiniated and old, so probably at least partly dated). If you can find it, the book "Tog on interface" gives some details about what Apple did to measure this (from what I remember, they moved the menu miles away from where the mouse was, and the menu still beat the keyboard due to Fitt's law)

http://ux.stackexchange.com/questions/30682/are-there-any-re... gives some more recent studies on the subject. I haven't read them, but it seems to me that keyboarding can be faster than mousing if you do it very, very, often.

And of course, the quality of mouse and keyboard matters a lot, too (early Windows mice were bad compared to Mac mice, mostly due to software (I have seen Windows mice were it was impossible to address every pixel on the screen, possibly due to staircasing/aliasing (what does one call that?) in the conversion between coordinate systems); this reminds me of http://www.folklore.org/StoryView.py?story=Shut_Up.txt, which isn't really related, but may be an interesting read, anyways)